Multiview Gait Recognition Based on Slack Allocation Generation Adversarial Network

Through the in-depth study of biological characteristics, gait information each person is found to be unique. Therefore, characteristics can used identify a person’s identity information. However, pedestrian often affected by various interference factors such as viewing angle, dress, and carrying objects. To solve this problem, idea perspective transformation proposed transform images from different perspectives states under standard conditions on one side. Generative adversarial network (GAN) adopted for transformation. Moreover, in order difficulty instability converge GAN training process, we introduced slack module adjust shortcomings GAN. Experiments show that normal walking conditions, our method achieves an average accuracy 98.32% CASIA-B dataset.